See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/45490123

Polyphenols in countercurrent chromatography. An example of large scale

separation

Article  in  Analusis · November 1999

DOI: 10.1051/analusis:1999140

CITATIONS READS

29 119

3 authors, including:

Alain Berthod Stephanie Geoffroy

Claude Bernard University Lyon 1 Pharmaron
303 PUBLICATIONS   10,114 CITATIONS    3 PUBLICATIONS   29 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Ultra fast separations, D-amino acids in biological systems, and new ionic liquids View project

Emeritus status View project

All content following this page was uploaded by Alain Berthod on 25 April 2014.

The user has requested enhancement of the downloaded file.

Analusis, 1999, 27, 750-757
© EDP Sciences, Wiley-VCH 1999

Polyphenols in countercurrent chromatography.

An example of large scale separation1
A. Berthod*, B. Billardello and S. Geoffroy

Laboratoire des Sciences Analytiques, Université de Lyon 1, CNRS UMR 5619,

69622 Villeurbanne Cedex, France

Abstract. Polyphenols are sometimes difficult to separate in classical liquid chromatography. Countercurrent chromatography uses a biphasic liquid sys-
tem to separate the components of a mixture. A centrifugal field allows to use a liquid stationary phase in an open tube. The phase density difference and
the centrifugal field are the only parameters allowing the equilibrium between the two liquid phases. The big advantage of the technique in preparative
separation is the dual-mode capability of CCC. The role of the phases can be switched during a run. The mobile phase becomes stationary and vice versa.
Then no injected material can be left in the machine. The large scale separations of a flavonoid mixture and two tannin samples are described: choice of
the biphasic liquid system, analytical study, scaling-up, injection protocol, two step separation in case of dual-mode use. It was possible to inject 26 g of
a tannin sample in one run.
Keywords. Countercurrent chromatography – polyphenol – preparative chromatography – flavonoid tannin.

Introduction The main problem in CCC is to retain the liquid stationary

Polar compounds are commonly found in vegetal extracts.
Saponins, cardenolides, iridoids, alkaloids, antibiotics, tan-
nins, flavonoids, quercetrins xanthones, lignans and other
polyphenols or sugar or glycoside derivatives are examples
of polar compounds very commonly found in vegetal
extracts and used as active principles for their pharmaco-
logical properties [1]. They are often difficult to separate by
HPLC and countercurrent chromatography can be a very
useful tool to purify them [2-3].
Countercurrent chromatography (CCC) is the separation
technique that uses a liquid stationary phase. The mobile
phase is also a liquid phase, i.e., a biphasic liquid system is
used. It is important to realize that in most CCC separations
there are absolutely no countercurrent flowing liquids. The
name of the technique was coined by Yoishiro Ito, its inven-
tor [4]. Publishing more than 250 articles using it, he made
the CCC acronym so related to the liquid-liquid centrifugal
partition chromatography technique that it is now accepted
worldwide [5].
The advantages of the use of a liquid stationary phase are:
• a simple retention mechanism (liquid-liquid partition);
• no irreversible solute adsorption;
• high loading capability;
• large range of polarity;
• little solute modification, degradation or denaturation;
• original selectivity;
• few pH limitations;
• low cost.
phase when the liquid mobile phase is pushed through it.
Centrifugal forces are used in all modern CCC apparatuses
[2, 3, 5, 6]. The liquid stationary phase is retained in chan-
nels by a constant centrifugal field in hydrostatic (HS) appa-
ratuses [7]. It is retained in coils of Teflon tubes by a vari-
able field in hydrodynamic (HD) apparatuses [6]. Another
problem is the efficiency reduced to a maximum of one plate
per channel for HS apparatuses and two to three plates per
coil turn for HD machines. The efficiency of CCC machines
ranges from 100 plates to 2000 plates depending on the geo-
metrical characteristic of the machine but also on the bipha-
sic liquid system used and on the operating conditions.
In this work, the separation of polyphenols of vegetal ori-
gin is described and the scaling-up of a tannin separation is
presented.
Experimental
CCC Apparatuses
Three different HD CCC apparatuses were used. They are
described in table I. The analytical machine was constructed
by the late SFCC company (Société Française de Chromato
Colonne) (Éragny, France) that discontinued its CCC
machine production in 1992. The two Kromaton, models 2
and 3, preparative machines were built by SEAB (64, rue
Pasteur, 94807 Villejuif, France).

1. This work was presented at the chromatographic symposium SEP’99 in Lyon (France), March 31-April 2, 1999.
*Correspondence and reprints.
Received May, 11, 1999; revised July, 6, 1999; accepted July, 20, 1999.

750
 Original articles

Table I. Instrumental characteristics of the hydrodynamic CCC machines.

Model characteristics volume ratio tube length tube diameter number of turns
mL β m mm

SFCC 2000 3 spools rotating 52 0.56 29 1.6 133

vertically 156 0.56 87 1.6 400

Kromaton 2 2 spools rotating 94 0.67 19 2.5 30

horizontally 1070 0.67 218 2.5 345

Kromaton 3 2 spools rotating 190 0.80 94 1.6 125

horizontally 396 0.80 197 1.6 260
981 0.80 200 2.5 260
1972 0.80 402 2.5 520

The SFCC machine can work with one or three spools. The spools of the Kromaton machines are coiled with different tubes. The β ratio is r/R, the ratio
of the spool diameter over the rotation diameter (distance between the rotor axis and the spool axis). All three machines have a temperature regulation
system.

Other hardware as the machine rotor is rotating. Unless otherwise indicated,

Only the column is special in CCC. The peripherical mate-
rials, pumps, valves, detectors and recorders or integrators
are classical LC hardware. A Shimadzu LC-10AS pump was
used for flow rates up to 9.9 mL/min. A Shimadzu LC8A
prep-pump was used when higher flow rates were desired.
A Shimadzu SPD-6A UV detector and a Cunow DDL 21
evaporative light scattering detector were used connected to
a Shimadzu C-R5A and a C-R6A integrator, respectively.
Touzard et Matignon (Dukert, Courtabœuf, France) was the
supplier. Thin layer chromatography was done with Merck
10 × 10 plates code 60-F254.
Solvents
Ethyl acetate, butanol and butanone (methyl ethyl ketone or
MEK), methyl isobutyl ketone (MIBK) were from SDS
(Peypin, France). Methanol and acetic acid were obtained
from Fluka (Sigma-Aldrich, St Quentin Fallavier, France).
They were used as received. The CCC relevant physico
chemical properties are listed in table II. Water was deion-
ized and distilled.
we used the following protocol to prepare a CCC “column”:
first the CCC machine is filled at a high flow rate with the
liquid phase chosen to be the stationary phase. 200 min
(more than 3 hours) were needed to fill the 2L machine at
10 mL/min. Then the rotor is started at the desired speed
and the mobile phase is pushed in the machine in the cor-
rect way: from head to tail (top to bottom) if the mobile
phase is the denser liquid or from tail to head if it is the
lighter phase [3, 6, 8]. The mobile phase equilibrates with
the stationary phase turns after turns displacing part of the
later. Only the stationary phase is seen exiting the machine.
It is collected in a graduated cylinder. When the mobile
phase exits out of the machine, the CCC column is ready.
The collected stationary phase volume corresponds to VM,
the volume of the mobile phase inside the machine that is
the dead volume. The stationary phase volume retained by
the machine is VS = VT – VM. These volumes are used to
calculate P, the liquid-liquid partition coefficient of an
injected solute using its retention volume, VR:
VR = VM + PVS = VT + (P - 1)VS (1)

Samples If the partition coefficient is known, then Eq. (1) allows to

predict accurately the retention volume of the solute using
The Quercetrin extract was obtained from L. Light & Co. the phase volumes inside the CCC “column.” Often the
Ltd., Colinbrode, England. The Hamamelis extract was fur- retention ability of a CCC machine is rated by the phase
nished by C.A. Erdelmeier of Willmar Schwabe GmbH, retention factor, Sf:
Karlsruhe, Germany. The tannin mixture was the product
number 48812, Tannic acid, of Fluka.
Sf = VS/VT (2)

Protocol
The Sf factor depends on the CCC machine and also on the
Since a centrifugal field is used to hold the liquid stationary biphasic liquid system used and the operating conditions [2,
phase in the CCC machine, a CCC “column” exists as long 6].

751
Original articles
Table II. Physico-chemical properties of the solvents used.
Solvent density viscosity
g/cm3 cP % w/w 20 °C
solvent in water water in solvent
1-butanol 0.810 2.95 7.8
butanone 0.805 0.43 24
ethyl acetate 0.901 0.45 8.7
methanol 0.791 0.55 4 4
water 0.998 1.0 4 4
δ is the Hildebrand parameter. Butanone is also called methyl ethyl ketone (MEK).
Choice of the liquid system
How to define solvent polarity
Solvent polarity is roughly related to the interactions
between the solvent molecules. If there are no interaction or
only Van der Waals interactions between the molecules of a
pure solvent, this solvent is considered as apolar, e.g., alka-
nes, silicon oils, perfluorinated solvents. If there are inter-
actions between the solvent molecules, then the solvent is
said to be polar. Different polarity degrees are possible.
Weak interactions (dipoles-induced dipoles) will correspond
to weakly polar. Strong interactions are found in very polar
solvents (hydrogen bonds, ionic interactions in room tem-
perature ionic liquids). At the moment, there are different
polarity scales such as the Hildebrand scale based on the
solubility parameter, δ, defined as the work necessary to sep-
arate two solvent molecules. In chromatography, the Snyder
polarity scale, based on the eluting power of the solvent in
thin layer chromatography (TLC), is commonly used. In
organic synthesis, the Reichardt scale is most often used. It
is related on the transition energy for the solvatochromic
absorption band of a pyridinium -N-phenoxyde betaine dye.
Table II lists the polarity value in the three scales for the
solvent used. In the Hildebrand and Reichardt scales, the
order is the same: water is the most polar solvent, then
methanol, butanol, butanone and ethyl acetate is the less
polar. In the Snyder scale, butanol is considered as the less
polar solvent. Its high viscosity reduces its eluting power in
TLC. Such differences between the different polarity scales
are very common [9].
Two or three solvent systems?
The two solvent biphasic liquid system is the most simple
and convenient to use in CCC. It is simple because the mix-
ing of the two solvents will produce the two liquid phases
that are one solvent saturated by the other. The butanone-
water system forms two liquid phases particularly useful in
CCC because of their high polarity. The aqueous denser
phase contains 76 % w/w water and 24 % w/w butanone
(28 % v/v). The upper organic phase contains 10 % w/w
Solubility Polarity
δ Snyder Reichardt
20.1 27.2 3.9 60.2
10 19.2 4.7 32.7
3.3 18.2 4.4 22.8
29.3 5.1 76.2
48.6 10.2 100
water and 90 % w/w butanone (92 % v/v). In CCC, the
phase that is used as the mobile phase is depleted more
rapidly than the other phase. There is no problem to prepare
more mobile phase. The mutual saturation of the butanone-
water system is fast.
Often it is not possible to find a biphasic system able to
perform the desired separation with only two solvents, then
three solvents are used: a good solvent for the sample, a sec-
ond solvent as good for the sample as possible and making
a second liquid phase with the first solvent and a third sol-
vent that partitions between the two phases. Ternary phase
diagrams are used [7, 10]. The third solvent allows a fine
tuning of the partition of the sample components between
the two phases. Given the number of possible combinations
of three solvents (and more), it can be difficult to find the
best biphasic liquid system. General methods to find rapidly
the liquid system to perform a good CCC separation were
exposed in the literature [11-13]. Figure 1 shows the ternary
mass diagram of the water (good solvent)-ethyl acetate (sec-
ond phase solvent)-butanol system that was used for the sep-
aration of tannin samples. Point A locates the composition
of the prepared mixture. Points B and C locate the compo-
sition of the upper organic and lower aqueous phase
obtained when the A mixture is well equilibrated. The BC
line is the tie-line corresponding to the A biphasic mixture.
Table III lists the corresponding w/w and v/v compositions.
Separation of a flavonoid sample
The sample called “Quercitrin extract” was fractionated
using the simple butanone-water biphasic system. A rapid
analysis using the small volume coil of the Kromaton 2
machine showed in 20 min at 2 mL/min (organic mobile
phase) that the fractionation of the sample was possible.
However, the small coil did not have a sufficient efficiency
to resolve fully the sample: only two peaks were obtained.
The large volume coil of the Kromaton 2 machine was
then filled with the butanone saturated aqueous stationary
phase in more than two hours at 8 mL/min. Next the rotor
was started at 400 RPM and the organic mobile phase was
752
 Original articles

Table III. Composition of the A, B and C points of Figure 1.

Figure 1 butanol ethyl acetate water* density

composition % w/w % v/v % w/w % v/v % w/w % v/v g/cm3

A global mixture 3.0 3.5 44.3 46.5 52.7 50 B and C phases

B upper phase 4.9 5.5 90.6 90.5 4.5 4 0.90
C lower phase 1.2 1.5 8.5 9.3 90.3 89.2 0.99

* in most experiments, water was actually a 1 % acetic acid solution (pH 2.8)

Table IV. Chromatographic parameters corresponding to Figure 2.

compound tr Vr N k Pbutanone/water Pwater/butanone

min mL plates

kaempferol 67 201 800 0.95 10 0.10

quercetin 86 258 700 1.52 6.2 0.16
kaempferol rhamnoglucoside 100 300 640 1.90 4.9 0.20
quercitrin 129 387 (600)* 2.75 3.4 0.29

* estimated value (peak off scale)

k: retention factor calculated as Pwater/butanone* VS/VM
Experimental CCC conditions described in Figure 2 caption

pumped into the rotating machine at 3 mL/min in the tail- 650 mL and in 4 hours. Figure 2 shows the CCC chro-
to-head direction. The stationary phase retention factor was matogram of a 0.33 g injection. The compounds were iden-
90.4 % (VM = 102 mL, VS = 968 mL, VT = 1070 mL). The tified using mass spectrometry as described in a recent arti-
flavonoid sample was separated in four components with cle [14].

Table IV lists the chromatographic parameters that can be

established from the CCC chromatogram. The retention vol-
umes give the butanone/water partition coefficients of the
compounds. These physico chemical parameters are precious
data in hydrophobic studies, in quantitative structure-
activity relationships (QSAR) and quantitative structure-
retention relationships (QSRR) in pharmacological or chro-
matographic studies, respectively. CCC is a very powerful
tool to determine accurately the liquid-liquid partition coef-
ficients of solutes [8, 9, 15].

Table IV and Figure 2 also show that CCC compensates

Fig. 1. The ternary mass diagram of the system water-ethyl acetate-
butanol at 20 oC. Hatched areas: monophasic compositions (not
usable in CCC). Composition A was prepared to perform the tan-
nin separation. It gives Compositions B and C whose formulas are
listed in Table III.
easily its low efficiency by a high selectivity power.
Kaempferol and quercetin as well as the rhamnoglucoside
derivative and quercitrin differ by only an oxygen atom in
the R1 position (Fig. 2). This oxygen atom produces a 60 %
and 45% increase of the Pwater/butanone coefficient, respec-
tively (Table IV). Since the k retention factor is directly pro-
portional to the P coefficient, the same change is observed
on the values. This produced a selectivity factor, α, of 1.6
between kaempferol and quercetin and 1.5 between the glu-
coside derivatives (α = k2/k1). Such selectivity factors are
high enough to produce an almost baseline resolution of the
compound with an efficiency of only about 600 plates.

753
Original articles

Fig. 2. CCC chromatogram of the quercetin sample. Mobile phase:

butanone 3 mL/min, tail to head. Stationary phase: aqueous.
VT = 1070 mL, VM = 102 mL, VS = 968 mL, 400 rpm.
Injection: volume 9 mL of a 37 g/L solution, mass 330 mg.
Detection: UV 330 nm.

Mass separation of tannins

Tannins are polar polyphenols of vegetal origins. They are

difficult to separate because in a vegetal extract numerous
isomers with a very similar molecular basis coexist [19, 20].
The use of a liquid-liquid system to fractionate tannin sam-
ples produces original selectivities and preparative capabili-
ties that were already used [3, 6-7, 16, 19, 20]. We present
here the separation of the tannin sample sold by Fluka with
the name “Tannic acid” product 48812 valid only for the
batch #365176/1 (Feb. 12, 1997). We bought a second
250 g pot (batch #395301/1 (May 5, 1998)) that gave dif-
ferent chromatograms. The commercial tannin mixture was
used to select the suitable liquid system and investigate the
loading capabilities of our hydrodynamic CCC machines.
Optimization using the low volume machine
The SFCC 2000 machine was used to search for the opti-
mal liquid system. Using and adapting the Margraff
approach [12], it was found that the composition A that sep-
arates in organic phase B and aqueous phase C (Fig. 1) of
the water-butanol-ethyl acetate system was able to partition
correctly the tannin sample. Table III lists the exact compo-
sitions of the phases.
Fig. 3. The dual-mode use of a CCC machine. Steps 1 and 2 are
done with the aqueous phase in the head to tail direction. Step 3
is done with the organic light phase in the tail to head direction.
Bottom: the corresponding chromatogram.
It was not possible to separate the sample directly. The
dual-mode was used. Figure 3 explains graphically this
unique way to operate a CCC machine using the liquid
nature of the stationary phase [7, 17]. Step 1 is the normal
way to use a chromatograph with a mobile, say aqueous, and
a stationary, say organic, phase. The aqueous denser phase
moves in the head to tail way. In Step 2, it is shown that
solutes 3 and 4 would need a huge mobile phase volume to
elute. They are hydrophobic with a high affinity for the
organic apolar phase. The answer to elute them is to switch
the phase role: the stationary phase becomes the mobile
phase and vice versa. Step 3 shows that solutes 3 and 4 are
easily eluted by the organic mobile phase in the opposite tail
to head way. Furthermore, the separation initiated in Step 2
is improved in Step 3 since 4 moves faster than 3 with the
organic mobile phase. The advantages of the dual mode use

754
 Original articles

were evidenced in the partition coefficient determination [8, Table V. Chromatographic parameters of the tannin separation pre-
15] and general uses of CCC [7, 18]. sented in Figure 4.
The tannin sample was fractionated in 12 peaks. 7 peaks
eluted in the head to tail way with an aqueous mobile phase peak tr VR P calculation
of composition C (Fig. 1). 5 more peaks eluted in the min mL
reversed tail to head way with the composition B organic
mobile phase (not shown by Figure 4, see Table V). The 1 40.0 80.1 0.08
paramount advantage of the dual mode use of CCC is that 2 0.81 258 0.22 P = (VR - VM)/VS
nothing can be left in the machine if the organic phase vol- 3 0.805 111.3 0.48 VM = 74 mL
ume, Vorg, pushed in the reversed tail to head way is higher 4 74.4 148.7 0.96 Vs = 78 mL
than: 5 89.6 179.2 1.35 VT = 152 mL
6 0.998 289.0 2.76 Sf = 52%
Vorg s Vaq VS/(Vaq – VM) (3)
7 193.6 387.2 4.02
in which the subscript aq, S, and M refer to the volume of Mode inversion after 215 min (not shown in Figure 4)
aqueous phase pushed in the head to tail direction after the
sample injection in Steps 1 and 2 (Fig. 3), the volume of 8 3.21 6.42 67.0
organic phase that was retained and stationary in Steps 1 and 9 5.48 10.96 39.2 P = Vaq / VR
2, and the volume of aqueous phase that was mobile in 10 9.31 18.62 23.1 Vaq = 430 mL
Steps 1 and 2, respectively. The VS volume of organic phase 11 23.8 47.6 9.0
in the machine should remain the same during Step 3. 12 37.8 75.6 5.7
Usually it does [7, 9, 15, 17, 18, 20].
Tannins are very sensitive to small pH changes, they oxi-
dize readily in basic media. To stabilize the aqueous phase
pH at a 2.8 value, 1 % v/v acetic acid was added to water prepared by dissolving the exactly weighted amount of tan-
before preparing the biphasic system. Table V lists the reten- nins in the adequate aqueous phase volume. A 300 g/L solu-
tion values and partition coefficients corresponding to the tion was prepared by dissolving 30 g of the tannin sample
peaks obtained. The partition coefficients listed correspond in a mixture of 5 mL of aqueous and 5 mL of organic phase.
to the affinity of the compounds for the organic phase The resulting solution was monophasic. Figure 4 shows the
(Porg/aq). The affinity for the aqueous phase would be mea- chromatogram obtained when 30 mg (1 mL of the 30 g/L
sured by the inverse: Paq/org = 1/Porg/aq. The peak elution solution, Fig. 4-A) and 750 mg (2.5 mL of the 300 g/L solu-
order from 1 to 7 (Fig. 4) corresponds to the increasing tion, Fig. 4-B) were injected in the 150 mL machine. The
Porg/aq order. The P values are obtained by Eq. 1. The mode loading capability of the CCC machine are demonstrated.
inversion is done after a volume Vaq of aqueous phase is Almost no broadening can be seen on Peaks #6 and #7
pumped in the head to tail direction. Then, the mobile phase between the 30 mg and 750 mg injections. Peaks 1, 4 and
becomes the organic phase pumped in the tail to head direc- 5 broaden somewhat, but they correspond to a twenty-time
tion (Fig. 3). The peaks elute in decreasing Paq/org order higher amount of tannins. The retention volumes are not
(Table V). The P value is calculated as P = Vaq/VR [15, 17]. affected by the concentration increase.
The volume of organic phase needed to elute all the
hydrophobic tannins out of the CCC machine is 95 mL The 300 g/L solution had an elevated viscosity. Injection
(Eq. 3). When 100 mL of organic phase are pumped in the loops with volumes higher than 2.5 mL could be made but
tail to head direction, it is certain that no injected compound a pressure increase due to the sample viscosity limited their
remains in the machine. This experiment shows that CCC use (1/16" tubing, 0.5 mm i.d.). The Kromaton CCC
was able to separate peaks corresponding to compounds with machine can use 1.6 mm and 2.5 mm i.d. tubing in its injec-
P values ranging from 0.08 (hydrophilic) to 67 (less polar). tion port. A 66 mL loop was made allowing to inject 20 g
However tannins are so complex that the separated peaks did when the 300 g/L solution was prepared. The 2L spool con-
not correspond to a single component. A thin layer chro- figuration was equilibrated with the Table III liquid system.
matography plate was done for every peak. All of them, The organic phase retention ratio was only 26% at 420 rpm.
except maybe peak #8 showed more than one spot after elu- The 66 mL injection of the viscous tannin solution produced
tion with an AcOEt-MIBK-Acetic acid (50-50-1 v-v-v) an important stationary phase leak. The injected plug forms
mobile phase and revelation with a FeCl3 pulverization (blue a piston pushing the biphasic liquid system out of the
spots). Evaporating the collected fraction, it was found that machine and no separation occurs. After some unsuccessful
Peak #8 contained more than 80% of the injected mass. injections, the protocol was modified as follows: 17 min
were needed to inject the sample at 4 mL/min, then 120 mL
Scaling-up the separation of aqueous mobile phase were further introduced in the
machine at 4 mL/min for 30 more minutes, next the flow
The tannin separation was scaled-up injecting larger rate was stopped, the mode valve was switched in the oppo-
amounts on the same machine. A 30 g/L solution was site way (tail to head) and the flow rate was resumed for

755
Original articles
Fig. 4. Separation of a vegetal tannin sample. A- separation of
30 mg (1 mL of 30 g/L solution) by the SFCC machine
(VT = 152 mL). B- Scaling-up with the same machine. Separation
of 0.75 g injected in 2.5 mL of a 300 g/L solution. 800 rpm, aque-
ous mobile phase in the head to tail direction, 2 mL/min, Sf = 56 %.
C- Large scale separation (20 g in 66 mL of a 300 g/L solution)
with the Kromaton 3 2L-preparative machine. 420 rpm, 4 mL/min,
Sf = 26 %. 280 nm UV detection. Liquid system: see Table III.
10 min. This was designed to dissolve the plug of injected
phase and to force the sample to partition between the two
liquid phases. After that, the flow rate was stopped again,
the machine was let to equilibrate for 5 min, the mode valve
was switched back in the correct position (head to tail) and
the flow rate was resumed at 4 mL/min. Figure 4-C shows
part of the obtained chromatogram. The retention volume of
Peak #7 was 3.3 L (almost 14 hours retention time).
The important noise observed on the chromatogram 4-C
is due to a significant stationary phase leak. Microdroplets
of stationary phase pass through the detector producing the
756
numerous spikes making the noise. The organic stationary
phase loss was measured to be 20 mL/h. The Sf phase reten-
tion ratio dropped from an initial 26 % value (VS = 515 mL)
to Sf = 11 % (VS ~210 mL) at the end of the head to tail
phase of the preparative tannin separation. The injection pro-
tocol and the separation done with a continuously decreas-
ing stationary phase volume modified somewhat the chro-
matogram aspect. It is still possible to note that the peak
looks gaussian with minimal distortion (Fig. 4-C). This was
not true during the tail to head phase of the separation. A
huge and broad peak was obtained, but it was not possible
to locate the 5 peaks obtained with the analytical CCC
machine. This is likely due to the stationary phase volume
reduction. The stationary phase loss could easily be reduced
or even stopped by increasing the centrifugal field (rotation
speed). The apparatus we used was a prototype limited to a
450 rpm maximum rotation speed. This rotation speed may
be not enough to create a centrifugal field song enough to
retain correctly polar biphasic liquid systems. The rotor mass
is in the 30 kg range when loaded with the liquids. The over-
all rotation diameter (rotor and spool) is in the half meter
range. It is obviously difficult to make a well-balanced rotor
of this size and mass. The small vibrations generated by the
minute out-of-balance may became dramatic and dangerous
when the rotation speed is increased. This mechanical prob-
lem is under investigation and it should be resolved since a
600 to 800 rpm speed (2 to 4 times higher centrifugal field)
seems desirable.
The protocol established with the commercial tannin sam-
ple was applied to a real sample, a Hamamelis extract, pro-
posed by Schwabe (Germany). This extract had been puri-
fied to refine the glucoside hamameli-tannin. It contained a
large majority of the desired tannin along with non desired
tannins. The analytical separation of the Hamamelis tannin
sample with the Table III biphasic liquid system showed
4 peaks: 3 impurities (25 % w/w) and a major peak con-
taining about 3/4 of the injected mass. The UV spectra of
the collected fractions were too similar to allow the charac-
terization of the hamameli-tannin. We suppose that the major
tannin was the desired one. The preparative separation with
the 2L Kromaton machine was done in two steps: 11 hours
(2.6 L) in the head to tail direction with the aqueous mobile
phase and 6 hours (1.4 L) in the opposite tail to head direc-
tion with the organic mobile phase. The major peak (more
than 70 % of the injected mass) was the first to elute with
a 390 min retention time or 1560 mL retention volume. If
this peak was Hamameli-tannin, it was completely eluted in
420 min (1.7 L). If the other tannins are not wanted, it is
possible to stop the CCC machine after 420 min, to empty
it rapidly using a compressed gas and to start a new purifi-
cation of 25 g. With these conditions, the throughput of the
production can be estimated to 20 g in 8 hours or 2.5 g/h.
Conclusion
The preparative capability of CCC is one more time demon-
strated. Large volume CCC machines that are appearing on
the market, are able to separate large mass of sample in one
 Original articles

run giving throughput comparable with classical prep-LC. 5. Ito, Y. “CCC” in J. Chromatogr. Library, Heftmann, E. Ed.,
The original selectivity obtained with the biphasic liquid Elsevier, Amsterdam, 1992, 51A, 69-105.
system chosen is associated with the paramount advantage 6. Conway, W.D. Countercurrent Chromatography, Apparatus,
of the guaranteed total recovery of the injected mass. Other Theory and Applications; VCH Publishers: Weinheim, 1989.
chromatographic modes not presented in this work, such as 7. Foucault, A.P. (Ed.) “Centrifugal Partition Chromatography”
displacement chromatography or pH zone refining [19], are Chromatographic Science Series 1995, 68.
able to separate even higher mass of sample in one run. As 8. Berthod, A. Spectra 2000 1989, 144, 34-38 and 1992, 169, 22-
the reliability of the CCC machines increases (and the noise 24.
they generate decreases !), they will be more and more used 9. Berthod, A. in Foucault, A.P. (Ed.) “Centrifugal Partition
as powerful tools to purify large mass of compounds. CCC Chromatography” Chromatographic Science Series, Chapter 7,
machines are already parts of production processes in the 1995, 68, 167-197.
pharmaceutical industry. 10. Berthod, A.; Duncan, J.D.; Armstrong, D.W. J. Liq.
Chromatogr. 1988, 11, 1171-1185.
11. Foucault, A.P.; Chevolot, L. J. Chromatogr. A 1998, 808, 3-
Acknowledgments 22.
12. Margraff, R. in Foucault, A.P. (Ed.) “Centrifugal Partition
François De La Poype, Director of the S.E.A.B, company is Chromatography” Chromatographic Science Series, Chapter 4,
gratefully acknowledged for the loan of a prototype of the 1995, 68, 80-87.
Kromaton 3 CCC apparatus. C.A.J. Erdelmeier of the 13. Oka, F.; Oka, H.; Ito,Y. J. Chromatogr. A 1991, 538, 99-111.
Schwabe GmbH company (Karlsruhe, Germany) is thanked 14. Berthod, A.; Talabardon, K.; De La Poype, F.; Erdelmeier,
for the gift of 80 g of a refined Hamamelis extract. C.A.J. J. Liq. Chromatogr. 1998, 21, 3003-3019.
15. Berthod, A.; Carda-Broch, S.; Alvarez-Coque, M.C.G. Anal.
Chem. 1999, 71, 879-888.
References 16. Marston, A.; Hostettman, K. J. Chromatogr. A 1994, 658, 315-
341.
1. Hostettmann, K.; Hostettmann, M.; Marston, A. Natural
17. Menges, R.; Bertrand, G.; Armstrong, D.W. J. Liq.
Products Reports 1984, 471-481.
Chromatogr. 1990, 13, 3061-3077.
2. Conway, W.D. Chapter 1 in Modern Countercurrent 18. Agnely, M.; Thiebaut, D. J. Chromatogr. A 1997, 790, 17-30.
Chromatography, ACS Symposium Series 1993, 593, 1-14.
19. Maillard, M.; Marston, A.; Hostettmann, K. in Ito, Y.; Conway,
3. Ito, Y.; Conway, W.D., “High Speed Countercurrent W.D., Chemical Analysis; J. Wiley: New York, 1996, 132, Ch.
Chromatography”, Chemical Analysis, J. Wiley, New York, 7, 179-223.
1996, 132. 20. Okuda, T; Yoshida, T.; Hatano, T.; Yazaki, K.; Kira, R.; Ikeda,
4. Ito, Y.; Bowman, R.L. Science 1970, 167, 281-283. Y. J. Chromatogr. 1986, 362, 375-389.

757

View publication stats